Direction giving and following 1 Running head: DIRECTION GIVING AND FOLLOWING Direction Giving and Following in the Service of Wayfinding in a Complex Indoor Environment Alycia M. Hund and Amanda J. Padgitt Illinois State University Direction giving and following 2 Abstract We examined how sense of direction, descriptive feature content, and gender relate to direction giving and following during wayfinding in a complex indoor environment. In Experiment 1, participants provided directions to destinations. Participants with a good sense of direction provided more distances, marginally more correct descriptions, and marginally fewer straight references than those with a poor sense of direction. In Experiment 2, participants rated the effectiveness of these directions. Directions that were rated highly contained more descriptive features than did directions that were rated less highly. In open-ended responses, positive mentions of landmarks and negative mentions of cardinal descriptors were frequent. In Experiment 3, participants navigated faster when following the worst-rated directions than when following the best-rated directions. Keywords: wayfinding, sense of direction, direction giving, direction following Direction giving and following 3 Direction Giving and Following in the Service of Wayfinding in a Complex Indoor Environment Finding our way through the environment is essential to human functioning. Often, people give and follow directions to facilitate wayfinding. For example, a college freshman may ask another student or a university staff member how to get to a classroom or office in a campus building. It is no surprise that spatial skills and strategies differ across individuals (e.g., Hegarty, Montello, Richardson, Ishikawa, & Lovelace, 2006; Kato & Takeuchi, 2003; Lawton, 1996; Prestopnik & Roskos-Ewoldsen, 2000). However, surprisingly little research has examined how these differences affect wayfinding involving direction giving and following in everyday environments. The primary goal of this study was to examine the processes involved in giving and following directions in the service of wayfinding in a complex indoor environment. In particular, we sought to specify how descriptive features, sense of direction, wayfinding strategies, and gender are related to direction giving and following for wayfindng in a university building. Descriptive Features in Wayfinding Directions People provide a variety of details when giving directions for wayfinding, including landmarks, cardinal directions, street names, distances, and turn descriptions (Golding, Graesser, & Hauselt, 1996; Hund, Haney, & Seanor, 2008; Mark & Gould, 1995; Ward, Newcombe, & Overton, 1986; Wright, Lickorish, Hull, & Ummelen, 1995). Individual differences in wayfinding details are widespread, with some people providing only the basic details, whereas others provide elaborate descriptions (Devlin, 2003). For instance, Denis, Pazzaglia, Cornoldi, and Bertolo (1999) found substantial differences in the length and amount of landmark information given when participants described three different routes in Venice. Direction giving and following 4 Given this variability, an important question is what factors impact the effectiveness of wayfinding directions. One important consideration is the features of the directions themselves. Allen (1997) describes these features in terms of environmental features (i.e., landmarks, pathways, choice points), delimiters (i.e., distance, cardinal direction, left-right), verbs of movement (i.e., turn, go, continue), and state-of-being verbs (i.e., you will be on x street, the destination is across from y). Some of these features are more preferred in route directions than others. For instance, Lovelace, Hegarty, and Montello (1999) asked participants to provide directions for familiar and unfamiliar routes. In general, landmark mentions correlated with route quality for both familiar and unfamiliar routes, indicating that higher quality routes contained more landmarks. Moreover, longer route descriptions received higher ratings because they were more complete. Furthermore, Denis et al. (1999) found that directions deemed high in quality were clear and complete, with an adequate number of landmarks, but with no redundancy or uncertainty. Poor-quality descriptions, on the other hand, were unclear, incomplete, or redundant (see also Devlin, 2003). Hund et al. (2008) asked participants to respond to an open-ended question about their preferences regarding wayfinding directions. Positive mentions of landmark and left-right information and negative mentions of cardinal directions were common, further supporting landmark preferences overall. In addition to assessing direction giving, it is important to specify direction following processes. For instance, Allen (2000) found that participants follow directions with landmarks at choice points with fewer errors than those with landmarks at non-choice points. Furthermore, directions containing landmarks led to fewer wayfinding errors than descriptions containing cardinal directions and distance information. Nonetheless, research findings regarding how effectively participants follow highly-rated or poorly-rated wayfinding directions are inconsistent Direction giving and following 5 (Denis et al., 1999; Honda & Nihei, 2004; Hund et al., 2008). In Denis et al.’s (1999) study, participants were asked to provide descriptions of three routes in Venice. These descriptions were compiled so that a new group of participants could rate their quality in navigational assistance on a seven-point scale. The highest and lowest rated descriptions then were included in a third experiment where participants were asked to follow the descriptions to the best of their ability. Participants navigated with fewer errors when following highly rated directions in comparison to poorly rated directions (see Lovelace et al., 1999 for similar ideas). In contrast, other research has shown that worst-rated directions facilitate faster wayfinding than best-rated directions. Hund et al. (2008) employed a similar methodology to Denis et al. (1999), except they had participants give descriptions of six routes in a model town. When participants followed the best- and worst-rated descriptions through the model town, they navigated faster following the worst rated directions. Why might this have been the case? It is possible that people navigated more quickly when following the worst-rated descriptions because these worst-rated descriptions were concise and to the point, which led to better wayfinding than the overly elaborate, specific best-rated descriptions. These inconsistencies across studies show that this area of research needs to be explored further. Moreover, it would be helpful to clarify what role environmental scale plays in these inconsistencies. Many different types of environments have been used to understand wayfinding processes, including virtual, model, indoor, and outdoor environments that differ in scale and mode of learning. For example, Hegarty et al. (2006) compared three different types of environments for wayfinding: a desktop virtual environment, walking two floors of a campus building, and watching a videotaped route through the building. Participants were made aware of certain landmarks on the route they learned and were asked to make distance and directional Direction giving and following 6 judgments between landmarks and to draw a sketch of the route including the landmarks. Although performance on direct learning measures differed from performance on measures derived from video or virtual learning, there were important relations among performance across domains. In particular, abilities related to small-scale space (e.g., embedded figures, mental rotation, spatial working memory, and perspective taking measures) were predictive of environmental learning, particularly learning through media. The authors therefore conclude that small- and large-scale spatial abilities rely on similar, but not identical, processes. In both Hund and Minarik (2006) and Hund et al. (2008), a table-top model town was used to assess wayfinding. Hund and Minarik (2006) found that cardinal descriptors led to higher wayfinding efficiency. Similarly, the worst-rated descriptions that led to better wayfinding performance (noted above) contained more cardinal descriptions than the best-rated descriptions (Hund et al., 2008). This pattern of results differs from the results of other studies that used larger, everyday environments (Denis et al., 1999; Saucier, Green, Leason, MacFadden, Bell, & Elias, 2002). Perhaps these variations are due in part to differences in interacting with small- and large-scale environments. In the model town, the environment was experienced via a survey perspective, perhaps rendering cardinal descriptions efficient for wayfinding. Furthermore, the entire environment was visible throughout the task, which reduced memory demands relative to everyday wayfinding in large-scale environments that involve ground-level views in which only part of the route is visible at any given moment. Although Hund et al. (2008) noted similarities in direction giving when conducting a direct comparison of small- and large-scale environments, potential similarities and differences between environments when following directions in the service of wayfinding need to be explored further. Sense of Direction Direction giving and following 7 Sense of direction, which is “an awareness of orientation or location” (Kozlowski